Analysis of Shear Wall Transfer Beam Structure
by
LEI KA HOU
Final Year Project report submitted in partial fulfillment
of the requirement of the Degree of
Bachelor of Science in Civil Engineering
2013-2014
Faculty of Science and Technology
University of Macau
DECLARATION
I declare that the project report here submitted is original except for the source
materials explicitly acknowledged and that this report as a whole or any part of this
report has not been previously and concurrently submitted for any other degree or
award at the University of Macau or other institutions.
I also acknowledge that I am aware of the Rules on Handling Student Academic
Dishonesty and the Regulations of the Student Discipline of the University of Macau.
Signature : ____________________________
Name : ____________________________
Student ID : ____________________________
Date : ____________________________
APPROVAL FOR SUBMISSION
The project report entitled “Analysis of shear wall-transfer beam structure”
was prepared by Lei Ka Hou in partial fulfillment of the requirements for the Degree
of Bachelor of Science in Civil and Engineering at the University of Macau.
Endorsed by,
Signature : ____________________________
Supervisor : Prof. Iu Vai Pan
I
ABSTRACT
The aim of this paper is to investigate the structural behavior and both the
stresses in the shear wall and in the transfer beam, which is usually placed at the first
floor to support the shear wall to make a large opening at ground level. This paper
presents an analysis and investigation of the structural behavior of transfer
beam-shear wall systems in tall buildings with different amount of span of shear wall
and geometry such as, span length, size of wall, beam and column. The different
between normal deep beam and the transfer beam which supporting shear wall is that
for normal deep beams, the estimation for structural behavior and failure mechanism
can be done by span/depth ratio table but not for that kind of transfer beam.
This paper mainly investigates the effects of stress in lower part of shear wall
and transfer beam due to different geometry and relevant parameters of the shear
wall-transfer beam system such as span/depth ratio, stiffness of the supporting
columns, and amount of span of the shear wall. All of models in this paper were
analyzed by computer software ABAQUS which is a structure analysis program based
on finite element method. The result has shown the significant effects of stress due to
the interaction of the different amount of span and geometry of shear wall, transfer
beam and column. Conclusions of the investigation show the structural behavior,
analysis and parametric studies.
II
ACKNOWLEDEMENT
The author would like to express his gratitude sincerely to the supervisor
Professor Iu Vai Pan for his valuable suggestions and kind supervisions. Besides, the
author would also like to thank family and friends for spiritual supporting.
CONTENTS
CHAPTER 1 INTRODUCTION ................................................................ 1
1.1 General ............................................................................................................................ 1
1.2 Objective of study ........................................................................................................... 2
1.3 Review of previous research ........................................................................................... 3
1.4 Scope of work .................................................................................................................. 8
1.5 Coordinate system and unit ........................................................................................... 10
1.6 Notation ......................................................................................................................... 10
CHAPTER 2 MODELING ....................................................................... 15
2.1 Introduction ................................................................................................................... 15
2.2 Finite element method ................................................................................................... 15
2.3 Structural modeling ....................................................................................................... 18
2.4 Meshing ......................................................................................................................... 20
CHAPTER 3 EFFECT OF SPAN ............................................................ 39
3.1 Introduction ................................................................................................................... 39
3.2 Structural behavior ........................................................................................................ 42
3.3 Results and analysis ....................................................................................................... 43
2.3 Summary ....................................................................................................................... 64
CHAPTER 4 EFFECT OF SPAN/DEPTH RATIO ................................. 68
4.1 Introduction ................................................................................................................... 68
4.2 Structural behavior ........................................................................................................ 72
4.3 Results and analysis ....................................................................................................... 73
4.4 Summary ....................................................................................................................... 88
CHAPTER 5 CONCLUSION .................................................................. 90
5.1 Effects on structural behavior ........................................................................................ 90
5.2 Recommendation for future study ................................................................................. 90
References ................................................................................................. 92
LIST OF FIGURE
FIGURE 1.1 77 WEST WACKER DRIVE ................................................................... 2
FIGURE 1.2 CASSELDEN PLACE ............................................................................. 2
FIGURE 1.3 DISTRIBUTION OF VERTICAL STRESS ............................................ 5
FIGURE 1.4 DISTRIBUTION OF HORIZONTAL STRESS ...................................... 6
FIGURE 1.5 DISTRIBUTION OF SHEAR STRESS ................................................... 6
FIGURE 1.6 TRIANGULAR TENSION ZONE .......................................................... 7
FIGURE 2.1 TYPICAL MESHED MODEL ............................................................... 20
FIGURE 2.2 SIMPLE MODEL EXAMPLE ............................................................... 23
FIGURE 2.3 TYPICAL SAMPLE PARTITION ......................................................... 24
FIGURE 2.4 MESHED MODEL (MODEL A-1) ........................................................ 26
FIGURE 2.5 MESHED MODEL (MODEL A-2) ........................................................ 27
FIGURE 2.6 MESHED MODEL (MODEL A-3) ........................................................ 28
FIGURE 2.7 MESHED MODEL (MODEL A-4) ........................................................ 29
FIGURE 2.8 MESHED MODEL (MODEL A-5) ........................................................ 30
FIGURE 2.9 CONVERGENCE OF VERTICAL STRESS AT POINT A ................... 32
FIGURE 2.10 CONVERGENCE OF VERTICAL STRESS AT POINT B ................. 32
FIGURE 2.11 CONVERGENCE OF VERTICAL STRESS AT POINT C ................. 33
FIGURE 2.12 CONVERGENCE OF HORIZONTAL STRESS AT POINT D .......... 33
FIGURE 2.13 CONVERGENCE OF HORIZONTAL STRESS AT POINT E ........... 34
FIGURE 2.14 CONVERGENCE OF HORIZONTAL STRESS AT POINT F ........... 34
FIGURE 3.1 MODELS WITH DIFFERENT SPAN (SET B MODELS) ................... 42
FIGURE 3.2 VERTICAL STRESS FOR SET B MODELS ........................................ 45
FIGURE 3.3 HORIZONTAL STRESS FOR SET B MODELS .................................. 49
FIGURE 3.4 HORIZONTAL STRESS ALONG DEPTH AT MID SPAN .................. 51
FIGURE 3.5 SHEAR STRESS FOR SET B MODELS .............................................. 55
FIGURE 3.6 VERTICAL STRESS FOR SET C MODELS ........................................ 58
FIGURE 3.7 HORIZONTAL STRESS FOR SET C MODELS .................................. 60
FIGURE 3.8 SHEAR STRESS FOR SET C MODELS .............................................. 63
FIGURE 4.1 ONE SPAN MODELS WITH SPAN/DEPTH RATIOS ........................ 70
FIGURE 4.2 TWO SPAN MODELS WITH SPAN/DEPTH RATIOS........................ 72
FIGURE 4.3 VERTICAL STRESS FOR SET D MODELS ....................................... 75
FIGURE 4.5 HORIZONTAL STRESS FOR SET D MODELS ................................. 78
FIGURE 4.6 SHEAR STRESS FOR SET D MODELS .............................................. 79
FIGURE 4.7 VERTICAL STRESS FOR SET E MODELS ........................................ 81
FIGURE 4.8 HORIZONTAL STRESS FOR SET E MODELS .................................. 83
FIGURE 4.9 SHEAR STRESS FOR SET E MODELS .............................................. 85
FIGURE 4.10 HORIZONTAL STRESS FOR MODELS ........................................... 87
1
CHAPTER 1 INTRODUCTION
1.1 General
Tall buildings emerged in the late nineteenth century in the United States of
America. Therefore, structural systems for tall building were rapidly developed from
that time. With different efficient height limit, the interior structures can be classified
as rigid frames, braced hinged frames, shear wall frames and outrigger structures.
Shear wall is a one type of lateral load resisting system usually used in tall buildings.
The effective height limit for using shear wall combined with rigid frame as lateral
load resisting system is around 70 meters.
The advantage of using shear wall is that it can resist lateral shear by concrete
shear wall effectively. The representative tall buildings that using shear wall as lateral
resisting system are 77 West Wacker Drive (Chicago, USA, 50stories, 203.6m) and
Casselden Place (Melbourne, Australia, 43 stories, 160m) which show at Fig. 1.1 and
Fig. 1.2 respectively. The disadvantage of using shear wall as a lateral system is that it
would limit the design of floor plan of ground floor if the shear wall is directly
supported by foundation. Therefore, shear wall are sometimes sitting on a deep
transfer beam, which takes all vertical load and lateral load from shear wall, then
spread to widely spaced big columns to vacate a large open space at ground floor.
However, the loading supported by transfer beam is very large and complicated
2
because much more parameters can affect the structural behavior due to the interaction
between shear wall and transfer beam.
1.2 Objective of study
The study concentrates on the behavior of shear wall-transfer beam structure for
tall building. This paper mainly investigates the behavior of interaction with various
independent parameters. The scopes of investigation are listed below:
Figure 1.2 Casselden Place Figure 1.1 77 West Wacker Drive
3
1. The structural behavior of shear wall in shear wall-transfer beam system while
considering the interaction between shear wall, transfer beam and columns and the
geometry of these elements.
2. The structural behavior of transfer beam in shear wall-transfer beam system with
different geometry of the system.
3. The effects of stress distribution due to the different geometries of beam and
columns.
1.3 Review of previous research
Shear wall is a very effective lateral resisting system. However, the existence of
shear wall usually affects the design of floor plan. Therefore, the shear wall-transfer
beam system has been investigated by numerous researches so that the larger openings
at ground floor can be achieved to give more space for architect to design the floor
plan. However, most researchers concentrated on the behavior of transfer beams and
shear wall in simple system. The geometry or relevant parameter of shear wall-transfer
beam system seems less important before. From resent year, the structural
development in tall building has been rapidly evolved. Therefore, the effect of
structural behavior due to geometry or relevant parameter of shear wall-transfer beam
system has been given more importance.
4
Jawaharlal, P. (1996) has investigated the behavior of shear wall-transfer beam
system. The results of paper have been cataloged in following:
1. In shear wall-transfer beam system, the region in the shear wall above the height
equal to L, the span length of transfer beam, from the wall-beam interface can be
considered as interactive zone. When shear wall subjects a distributed in-plane
load toward to the transfer beam, the vertical stress distribution at the interactive
zone present as an “arch” shape along horizontal direction. There are two kinds
of arches which are primary arch and secondary arch which are responsible to
transfer the major part of load to exterior support regions and transfer the
remaining part of load to exterior and interior support regions respectively. Fig.
1.3 shows the primary arch and secondary arch graphically. For a single span
shear wall-transfer beam system, only the primary arch effect occurs and for the
two span case of shear wall-transfer beam system in paper, both primary and
secondary arch effect occurs and the vertical stress at support can be increase to
two times of original applied load.
5
Figure 1.3 Distribution of vertical stress
2. Within the interactive zone which is just mentioned above, there is special effect
for horizontal stress, vertical stress and also shear stress. However, areas in shear
wall which above the interactive zone, the effects of these stresses disappear
rapidly. For elements which are not in the interactive zone, the vertical stress
distribution is the same as the load added on the top of shear wall. Moreover, The
horizontal stress and shear stress are approximately equal to zero, which means
there is no horizontal and shear interaction for those elements which out of the
interactive zone. Fig. 1.4 and Fig. 1.5 show the distribution of horizontal stress
and shear stress distribution in shear wall respectively.
6
Figure 1.4 Distribution of horizontal stress
Figure 1.5 Distribution of shear stress
7
3. Shear wall is always in compression in both vertical direction and horizontal
direction when it subjects an in-plane load toward to transfer beam. However, for
example in two span case of shear wall-transfer beam system, there is a tension
zone in the shear wall above the interior column which is approximately
triangular in area. The base and height of the triangular tension zone are
approximately 1.8-2.4 hc. Within this zone, the horizontal stress is positive and
the lower element in position, the larger tensile stress occurs. The tensile zone
can be recognized in Fig. 1.6 graphically.
Figure 1.6 Triangular tension zone
8
1.4 Scope of work
After the literature review, some of basic behaviors in shear wall have been
shown like the arch shape effect on the vertical stress of the lower part of shear wall
and the triangular tensile zone at shear wall above the interior support. Therefore,
there are many interesting studies that can be investigated. The mainly scope of work
in this paper is to present the structural behavior and analysis of different geometry of
shear wall-transfer beam system so that some kind of table like span/depth ratio can be
built up which can be used to estimate the geometry parameters of shear wall-transfer
beam system in different situation. Analysis of different considerations is separated in
several chapters. The following is the brief introduction about each chapter.
In Chapter 2, the modeling concept is discussed. For this paper, the method for
investigation is based on finite element method. Therefore, the concept of finite
element would be discussed. Besides, in order to obtain results in using ABAQUS,
there are series of steps and some of assumption which is needed to construct a model.
Therefore, the steps to construct a shear wall-transfer beam system in ABAQUS are
also discussed. Moreover, to obtain results using finite element program, there is one
step called “meshing” which is one of step in ABABUS, is to divide each partition to
be numerous elements to perform the finite element equations to each element. The
accuracy of results depends on the size of element. However, the size of elements
9
should not be too small because it is not efficiency. Therefore, the size of elements is
investigated which can obtain an accurate results in effective manner.
In Chapter 3, there are divided in to two parts. The first part is that the total
length of in-plane loaded shear wall and transfer beam is fixed. The arching action
over the supports can be highlighted with different span and span length. The vertical
stress and shear stress distributions at certain depths are plotted to clear the parameters
which affect interaction of shear wall-transfer beam system. The second part is that
the total length of span is fixed to see the stresses distribution when there is different
amount of span. Therefore, the behavior of stresses obtained from these two parts can
be compared to get more detail about the stresses effect due to parameters of span.
For normal beam, the estimation of size of beam can be checked by span depth
ratio table. However, these tables are not available for the transfer beam in shear
wall-transfer beam system because for the transfer beam supporting in-plane loaded
shear-wall, the interaction of shear wall and transfer beam would affect the result. In
chapter 4, the effect of stresses due to different depth of beam is investigated while the
span length is fixed.
The overall conclusion of this paper is discussed in Chapter 5.
10
1.5 Coordinate system and unit
In this paper, the coordinate system of all models is set. As all models are 2-D
problem, therefore the coordinate for models can just define in two directions. To
match the global axis used as usual, the horizontal axis set as x-axis and the vertical
axis set as y-axis.
The unit for this paper is in International system (SI unit). The unit for length is
meter and the unit for force is kN. Therefore, for the distributed in-plane load adding
on the top of shear wall in shear wall-transfer beam system, the unit is kN per meter.
Besides, the finite element program ABAQUS which is a dimensionless program.
Therefore, it should be cautioned that the unit of all data must be the same when
modeling. For all of stresses obtained from ABAQUS, the unit is kN per meter square.
1.6 Notation
In this paper, there are some of parameters will be used to obtain a better and
reasonable comparison on different stresses. These parameters are introduced in this
section.
11
1.6.1 Geometric and load parameters
There are some geometric and load parameters that would used in this paper.
Below shows the definition of those parameters:
L: Length of span
d: Depth of transfer beam
t: Thickness of shear wall
w: Distributed in-plane load per unit length
P: Pressure
1.6.2 Vertical stress
The vertical stress is the stress that is induced by external force in the vertical
direction. In finite element modeling, there is a vertical stress at each node which can
show the variation of vertical stress at any point. Besides, it can also show the
behavior of the model at a certain point. If the vertical stress is shows in positive value,
which means the element at that location is subjected to a compression force.
12
1.6.3 Horizontal stress
The horizontal stress is the stress that is induced by external force in the
horizontal direction. In finite element modeling, there is a horizontal stress at each
node which can show the variation of horizontal stress at any point. Besides, it can
also show the behavior of the model at a certain point. If the horizontal stress is shows
in positive value, which means the element at that location is subjected to a
compression force.
1.6.4 Shear stress
The shear stress is the stress that is induced by external force between two
elements. In finite element modeling, there is a shear stress at each interface of two
elements which can show the variation of shear stress at any interface.
1.6.5 Vertical stress parameter
In order to obtain the reasonable results and ratio to present the vertical stress
distribution at a level, the dimensionless ratio, which is the vertical stress
parameter developed to show the vertical stress variation. Below shows the equation
of the vertical stress parameter:
(2.1)
13
where
: The vertical stress parameter
The vertical stress
t: The thickness of shear wall
w: The distributed in-plane load per unit length
1.6.6 Horizontal stress parameter
In order to obtain the reasonable results and ratio to present the horizontal stress
distribution at a level, the dimensionless ratio, which is the horizontal stress
parameter developed to show the horizontal stress variation. Below shows the
equation of the horizontal stress parameter:
(2.1)
where
: The horizontal stress parameter
The horizontal stress
t: The thickness of shear wall
w: The distributed in-plane load per unit length
14
1.6.7 Shear stress parameter
In order to obtain the reasonable results and ratio to present the shear stress
distribution at a level, the dimensionless ratio, which is the shear stress parameter
developed to show the shear stress variation. Below shows the equation of the shear
stress parameter:
(2.1)
where
: The horizontal stress parameter
The horizontal stress
t: The thickness of shear wall
w: The distributed in-plane load per unit length
15
CHAPTER 2 MODELING
2.1 Introduction
Shear wall-transfer beam system is not like normal beam that can estimate the
size and behavior by span/depth ratio. The structural behavior of shear wall-transfer
beam system is sometimes hard to predict because the interaction between shear wall,
transfer beam and column is very complex. Therefore, the powerful structural analysis
method is used in this paper to perform the structural analysis which is finite element
method. ABAQUS is one of structural analysis software which based on finite element
method. The computation is performed with the finite element code ABAQUS in this
paper to generate the interactive forces and stresses at different locations for all
models. The axis used in this paper is global axis which presents horizontal direction
as x-axis and vertical direction as y-axis.
2.2 Finite element method
Finite element method is a numerical method to solve engineering problems and
mathematical physics. By this method, the approximate solution can be obtained but
rather than analytical solutions. This method is useful for problems with complicated
geometries, material properties and loadings where the solutions cannot be obtained
analytically. The purpose of using finite element analysis rather than analytical
16
solution is that finite element analysis can predict the performance and behavior of the
design, and to identify the safety margin accurately. Moreover, the physical behaviors
of a complex object can be understood. But for analytical solution the factor of safety
is usually given by experience.
There are six steps involved in finite element method which are element
discretization, primary variable approximation, element equations, Global equations,
Boundary conditions and solving the global equations sequentially. The introduction
about each step is presented following:
1. Element discretization
In this process, the geometry of investigating model is separated in many small
regions, called finite elements. There are nodes defined on each element, or within the
elements.
2. Primary variable approximation
In this step, a primary variable must be selected or assigned like a displacement or
load adding on any point of the model. This variable is expressed in terms of nodal
values.
3. Element equations
In this step, the element equation is generated as showing below:
(2.1)
17
where [KE], is the element stiffness matrix which include those parameters about the
material properties. , is the vector of incremental element nodal displacements.
is the vector of incremental element nodal forces. There is one element
equation for each element.
4. Global equation
In this step, element equations are combined together to form global equations shown
as below:
(2.2)
where [KG], is the global stiffness matrix. , is the vector of all incremental
nodal displacements. is the vector of all incremental nodal forces.
5. Boundary condition
In this step, the boundary conditions are formulated and the global equations are
modified. If there is loading applied on the model, the would be affected and
if there is displacements on the model, the would be affected.
6. Solving the global equations
The global equations can be solved after step 5. Then can be obtained which
is included the displacements at all the nodes. After is obtained, the secondary
quantities such as stresses and strains can be calculated.
18
There are three main groups of finite elements which are 1-D element, 2-D
element and 3-D element. In this project, 2-D element is used in analyzing models.
Therefore, the following will be just focused on introducing 2-D element.
There are two branches in analyzing 2-D problem which are plane stress analysis
and plane strain analysis. Plane stress analysis is used in problems such as a plate with
some changes or with holes in geometry that are loaded in plane. Plane stress analysis
is usually used in one of dimension of the model much smaller than the others while
the load is applied on the in-plane direction. Therefore, the stress perpendicular to the
x-y plane (the 2-D plane) is assumed to be zero. Plane strain analysis is used in
problems such as the geometry of model in one direction is much larger than others.
The load applied in x-y plane does not affect the z direction. Therefore, the strain
normal to the x-y plane, and the shear strain in x-z plane and y-z plane are assumed to
be zero. In this project, because the in-plane uniform distributed load is applied at the
top of shear wall, the plane stress analysis is used to investigate the structural
behaviors of the shear wall-transfer beam system.
2.3 Structural modeling
To simulate the shear wall-transfer beam system in ABAQUS, 2D planar
modeling space is used because the size of thickness of those parts is relatively much
19
smaller than the total length and the height of whole structure. The deformable type
and shell base feature are chosen to simulate the planar shell parts which are used to
model both of shear wall, transfer beam and columns. Moreover, the sections of all
parts are used homogeneous solid section to simulate the behavior of in-plane loaded
shear wall-transfer beam system, thus to ensure that there is no out of plane effect
occurs. The tie constraints are used in the interfaces between shear wall and transfer
beam, and also between transfer beam and columns to perform a perfect bonded in
those interfaces so that there is no displacement occurs due to sliding between parts.
The pin boundary condition is used in the bottom surfaces of columns. The mesh size
depends on the size of model in different Chapters; the detail meshing size is
discussed in Section 2.4. The plane stress element CPS4 is used in all parts. Fig. 2.1
shows the typical model which is investigated in this paper.
20
Figure 2.1 Typical meshed model
2.4 Meshing
2.4.1 Introduction
When the model is investigated by finite element method, meshing is necessary
for a model to separate many small elements then a series of equations shown in
Section 2.2 can be performed. The smaller elements in model meshed, the more
accurate result can be obtained. However, the judgment of the number of each element
should also be concerned because there is a certain number of each partition which
can get an accurate result sufficiently and smaller element used in analysis request
21
longer time and computing resources needed to obtain result. Therefore, a further
smaller element is not necessary for the model to obtain results.
2.4.2 Meshing consideration
There are several considerations when deciding the number of element. Below
shows some considerations in this project:
The size of model
The stress variation
The size of adjacent partition
1. The size of model
This is one of consideration when meshing a model. For 2-D problem, there are
two dimensions of length which should be meshed before performing analysis to
obtain result for a model. Therefore, if the length is larger, there should be more
elements in that dimension. Moreover, the size of element should be smaller if
the problem is just mainly on one of dimension in investigation.
2. The stress variation
For most of problems, the stress may increase or decrease rapidly like near the
point where adding a concentrated load or the regions near the column support.
22
Usually near those regions, the stress will increase rapidly because there are
some boundary conditions or stiffer region. Therefore, the smaller element
should be used in those regions to obtain an accurate result.
3. The size of adjacent partition
In order to get an accurate result through finite element method, the meshing of
each adjacent partition should be consistent. That means both of two dimensions
of the size of elements should be the same with the adjacent element. Therefore,
the common node can be formed between elements and the result will be more
accurate due to the deformation at common node is the same for those elements.
2.4.3 Modeling
In this section, the size of element request to obtain an accurate result is
presented. Considering a simple example that is a one span shear wall transfer beam
structure. The model is shown in Fig. 2.2 which shows three parts of partitions in
forming a shear wall-transfer beam model.
23
Figure 2.2 Simple model example
For the considerations listed above, the main investigating parts of a model are
the lower part of shear wall and the transfer beam. Therefore, the shear wall and the
transfer beam are chopped into several parts respectively. The separation of each
partition is shown in Fig. 2.3.
24
Figure 2.3 Typical sample partition
25
In Fig. 2.3, there are 1 to 5 parts. Part 1 and 2 are the upper part and the lower
part of the shear wall respectively; part 3 is the transfer beam. Part 4 and 5 are two
columns. In this paper, the main investigation partitions are the transfer beam and the
lower part of shear wall. Therefore, at these two partitions the size of elements is
defined as smaller than others to obtain an accurate result.
For the transfer beam, the horizontal stress distribution along the span will be
investigated. Therefore, the elements in horizontal direction should be defined as
smaller. Besides, the stress distribution along the transfer beam at mid span will also
be investigated. Therefore, the elements in vertical direction should also be defined as
smaller. Moreover, the stress variation can be changed rapidly near the supporting
columns. Therefore, the size of element in transfer beam should be smaller at those
regions.
For the shear wall, the elements in the upper part of shear wall can be reduced
because this is not the investigation part. However, for the lower part of shear wall,
the vertical stress and shear stress along the span will be investigated. Therefore, the
elements in horizontal direction should be defined as smaller.
There is a set of model which containing five different meshed models in this
section to investigate the convergence of the result due to meshing, named model set
A. For each model, six points are taken out as reference point to compare the
26
difference between each model named as point A, B, C, D, E and F which are shown
in Fig. 2.3. The meshing of models is shown in Fig. 2.4, Fig. 2.5, Fig. 2.6, Fig. 2.7 and
Fig. 2.8. In these figures, the number shown is the number of element.
Figure 2.4 meshed model (model A-1)
27
Figure 2.5 meshed model (model A-2)
28
Figure 2.6 meshed model (model A-3)
29
Figure 2.7 meshed model (model A-4)
30
Figure 2.8 meshed model (model A-5)
31
2.4.4 Method of comparison
After those models are analyzed by ABAQUS, the magnitude of stress at any
point of the model can be known. To investigate the convergence of the stress due to
the size of meshing, the comparison of the stress of each model at the six points
mentioned above should be performed.
2.4.5 Stress at reference point
There are six reference points for each model to compare the stress variation due
to the size of element. However, the vertical stress is the main concern in the lower
part of shear wall because it can show the arch effect due to the interaction between
shear wall and transfer beam. Therefore, the comparison of stress at point A, B and C
for each model is based on the vertical stress comparison. On the other hand, the main
concern in the transfer beam is the horizontal stress. Therefore, the comparison of
stress at point D, E and F for each model is based on the horizontal stress comparison.
2.4.6 Result
In this section, a series of figures are plotted to show the difference of stress
between models at each point. For each figure, the x-axis presents the number of
element used in the partition at that direction of investigating stress. The y-axis
32
presents the ratio of stress, to show the variation of stress. Fig. 2.9 to Fig. 2.14
shows the stress ratio results.
Figure 2.9 Convergence of vertical stress at point A
Figure 2.10 Convergence of vertical stress at point B
0
20
40
60
80
100
120
20 30 40 50 60 70 80
σy
Number of element
430
440
450
460
470
480
490
500
510
20 30 40 50 60 70 80
σy
Number of element
33
Figure 2.11 Convergence of vertical stress at point C
Figure 2.12 Convergence of horizontal stress at point D
0
50
100
150
200
250
300
20 30 40 50 60 70 80
σy
Number of element
0
5
10
15
20
25
30
35
15 25 35 45 55 65 75 85
σx
Number of element
34
Figure 2.13 Convergence of horizontal stress at point E
Figure 2.14 Convergence of horizontal stress at point F
21.5
22
22.5
23
23.5
24
24.5
25
25.5
26
26.5
15 25 35 45 55 65 75 85
σx
Number of element
123.8
124
124.2
124.4
124.6
124.8
125
125.2
15 25 35 45 55 65 75 85
σx
Number of element
35
From Fig. 2.9, it shows the vertical stress difference between different models at
point A. The result shows there is no different of the vertical stress when the size of
element becomes smaller and the vertical stress is remain the same as applied in-plane
load on the top of the shear wall. Therefore, the size of element at the upper part of
shear wall can be remained the same as those model.
From Fig. 2.10, it shows the vertical stress difference between different models at
point B. The figure shows the vertical stress variation is a little bit large compare with
model 1 and model 5 because the vertical stresses at point B are 440 and 503 for
model 1 and model 5 respectively. However, the figure shows that the vertical stress
obtained from model 3 and model 4 are also tended to constant. Therefore, the size of
element for the lower part of shear wall in vertical direction can use the same as model
A-3 or model A-4.
From Fig. 2.11, it shows the vertical stress difference between different models at
point C. The figure shows that the vertical stress tends to constant from the model A-2.
Compare with the Fig. 2.10, which shows the size of the element at the lower part of
shear wall should be the same as model A-3 or model A-4. Therefore, the size of
element for the lower part of shear wall can be the same as model A-3.
36
From Fig. 2.12, it shows the horizontal stress difference between different
models at point D. The figure shows that the horizontal stress tends to constant slowly.
The horizontal stress obtained from model A-4 and model A-5 is 30 and 32
respectively where the different is acceptable. However, the horizontal stress is very
small at this point compare to the stress obtained from the other points. Therefore, the
size of element at this partition can be the same at model A-4.
From Fig. 2.13, it shows the horizontal stress difference between different
models at point E. The figure shows that the horizontal stress tends to constant when
the size of element becomes the same as model A-4. Compare with Fig. 2.12, which
shows the size of element should also be the same as model A-4. Therefore, the size of
element for the transfer beam in horizontal direction can be taken as the same as
model A-4.
From Fig. 2.14, it shows the horizontal stress difference between different
models at point F. The figure shows that from the stress obtained from model A-2, the
stress keep consistent when the size of element becomes small. Therefore, the size of
element can be remained the same as model A-2.
37
2.4.7 Conclusion
In this section, the effect of stresses due to the size of element is investigated for
the sample model. For point A, the result shows that whatever the size of element is,
the vertical stress at this point will not change simply because at that level, the vertical
stress is still constant which means no arch effect on that level of shear wall. This is
also evidenced by the previous paper. Therefore, the number of elements in the upper
part of shear wall can be just the same as those models. For point B, D and E, the
stress variation for models changed slowly because the stress variation is large at these
points which are at the corner of the model. Especially for point E, that is the interior
point at the corner of the column support and the transfer beam. For point C and F, the
result shows the stress becomes constant quickly which is because at these points, the
variation of stress is not that large compare to point B, D and E. Therefore, for the
lower part of shear wall, the number of element should be used the same as model A-3.
For the part of transfer beam which supporting by column, the size of element should
be the same as model A-4 and for the interior part of transfer beam, the number of
element should be the same as model A-3.
In order to obtain the accurate result of stress for models with any geometry for
each partition from ABAQUS, the size of element for each partition should be
calculated. With the geometry of model shown in Fig. 2.3 and the results discussed
38
above, the size of element for each partition, which can obtain an accurate stress result,
can be calculated. Table 2.1 shows the number of element which should be used to
obtain an accurate result when model is be investigated.
Table 2.1 number of element for each partition
Partition Number of element in x
direction
Number of element in y
direction
Upper part of shear wall 58 20
Lower part of shear wall 58 50
Transfer beam 58 6
Column 4 4
39
CHAPTER 3 EFFECT OF SPAN
3.1 Introduction
The main reason to use a transfer beam to support shear wall is that it can take all
of vertical load and lateral load from shear wall and spread to columns. Therefore, the
larger space opening can be obtained at ground floor. However, the space between
columns is one of significant parameter that affects the structural behavior and stresses
in lower part of structure. Besides, the interaction effects from the interior column or
columns support have a significant effect on the structural behavior of shear wall like
primary and secondary arch. With the different criteria or constraint of the length of
shear wall, the length of each span is one of concern that affects the structural
behavior of shear wall significantly. Moreover, the vertical stress can also increase
significantly with different length of span.
In this Chapter, there are two different parts to investigate the effect of span for
the shear wall-transfer beam system. The first part is that models investigated with a
constant building width but different amount of span, the other part is that models
investigated with a constant length of span and with different amount of span. There
are two sets of models and for each set of model, there are three models which will be
investigated in each part. Table 3.1 shows the length of span of each model for
different parts.
40
Table 3.1 Span length of models
Span length (m)
Model 1 Model 2 Model 3
Total length fixed 15 7.5 5
Span length fixed 15 15 15
The geometry of models is shown in Fig. 3.1. These three models are simulated
when the top of shear wall subject an in-plane pressure P toward to transfer beam
where P equal to 100kN/m2. Therefore, with the thickness of shear wall 0.4m, the
distributed in-plane load w equal to 40kN/m.
41
(a) Model with one span (Model B-1)
(b) Model with two span (Model B-2)
42
(c) Model with three span (Model B-3)
Figure 3.1 Models with different span (Set B models)
3.2 Structural behavior
The structural behavior of a shear wall-transfer beam system subjected to vertical
loading is very complex due to the interaction of shear wall and transfer beam.
Besides, the interaction of exterior column and interior column is also a significant
role in stress analysis. These three models can simulate the structural behavior in one
span, two span and three span. Therefore, the shear wall-transfer beam system with
more span can be also estimated based on these models. The details behavior of
vertical stress in the lower part of shear wall can be investigated. Therefore, the
primary arch effect and secondary arch effect can also be investigated. However, from
43
previous research, the horizontal stress in shear wall is always in compression except
the tensile zone which is above the interior support. Therefore, the location of detail
investigation for horizontal stress is in the transfer beam because the horizontal stress
in transfer beam is much larger than the horizontal stress in the tensile zone. The
location of detail investigation for shear stress is in lower part of shear wall also
because the shear stress in the lower part of shear wall is always larger than the shear
stress which in the transfer beam. The positive value in stress means compression
stress and negative value means tension.
3.3 Results and analysis
3.3.1 Constant building width
The lengths of span for each model have been presented previous. For this
section, the total length of structure is fixed as 15 m. The length of span for two span
and three span is divided equally which is 7.5m and 5m respectively.
1. Vertical stress
In this section, the vertical stress distribution in the lower part of shear wall is
investigated with different length of span but the fixed total length. The three models
shown in Fig. 3.1 have been analyzed and for each model, the data of vertical stress
have taken from three horizontal sections where are y/L=0, y/L=0.5 and y/L=1 (L is
44
the span length of the model, in these three models, L=15m and y is the height of
shear wall counted from the bottom of shear wall) to see the variation of vertical stress
distribution with at these three different locations which is called the arch effect due to
the span length and at the different depth of shear wall in shear wall-transfer beam
system. These three levels are named as level S1, level S2 and level S3 respectively.
The variation of vertical stress distributions along the height of shear wall of
these three models are shown in Fig. 3.2.
(a) Variation of vertical stress along the shear wall (Model B-1)
-1
0
1
2
3
4
5
6
0 0.25 0.5 0.75 1
α
x/L
level S1
level S2
level S3
45
(b) Variation of vertical stress along the shear wall (Model B-2)
(c) Variation of vertical stress along the shear wall (Model B-3)
Figure 3.2 Vertical stress for set B models
0
0.5
1
1.5
2
2.5
3
3.5
0 0.25 0.5 0.75 1
α
x/L
level S1
level S2
level S3
0
0.5
1
1.5
2
2.5
0 0.25 0.5 0.75 1
α
x/L
level S1
level S2
level S3
46
For Fig. 3.2, it shows the vertical stress distribution of each model along the span.
For the results of these three models shown in Fig. 3.2, the largest vertical stress
occurs at the support of the end span. The smallest value of vertical stress occurs at
each mid span of the model. For the interior support, the vertical stress goes up again
but it is not very large compared to the end support. At the bottom of the shear wall,
the vertical stress at interior support is even less than half of the vertical stress occurs
at the end support.
For the model B-1, the vertical stress at end support is 5 times of the applied
pressure which is also the largest in these three models. As the length of span becomes
shorter, the vertical stress at end support reduced. Moreover, for model B-1 the
vertical stress becomes to zero at mid span but when the length of span becomes
shorter, the vertical stress at mid span goes up. That means the arch effect reduced
when the length of span becomes shorter. This can also be seen at the other level of
shear wall. For model B-1, the arch effect at level S2 can also been seen. However, the
arch effect almost disappears and the vertical stress tends to constant at that level for
model B-2 and model B-3.
The present of the arch effect is because the stiffer region at the column support.
When there is a column supporting a transfer beam, the column will provide a
constraint to the transfer beam to reduce the vertical displacement, which means stiffer
47
of that region as result and the vertical stress will becomes larger than those regions
that no constraint. For the interior column, the vertical stress is less than the exterior
column, which is because for the interior column, the left span and right span give an
opposite moment to each other. Therefore, the vertical stress can be reduced due to the
opposite moments. However, for the exterior column there is no another side of span
to balance the vertical stress. As result, the larger vertical stress occurs at the exterior
column.
2. Horizontal stress
For the investigation of horizontal stress in shear wall, it is always in
compression except the tensile zone which is likely a triangular area just above
interior supports. However, for horizontal stress investigation, the interest of
horizontal stress is more concentrate on the internal area of the transfer beam to see
whether the transfer beam is designed as fully tension beam due to the effect of shear
wall-transfer beam system or as an ordinary beam with the upper part subjected to
compression and the lower part subjected to tension. In this section, for each model,
there are three levels of depth of the transfer beam have been investigated where are
the bottom level, middle level and the top level of the transfer beam named as level T1,
level T2 and level T3 respectively. That is, the top surface, middle line and the bottom
48
surface of transfer beam are investigated to show the variation of horizontal stress
distribution with the different depth of the transfer beam.
The variation of horizontal stress distributions along the different height of
transfer beam of these three models are shown in Fig. 3.3.
(a) Variation of horizontal stress along the transfer beam (Model B-1)
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
0 0.25 0.5 0.75 1
β
x/L
level T1
level T2
level T3
49
(b) Variation of horizontal stress along the transfer beam (Model B-2)
(c) Variation of horizontal stress along the transfer beam (Model B-3)
Figure 3.3 Horizontal stress for set B models
-1.5
-1
-0.5
0
0.5
1
1.5
2
0 0.25 0.5 0.75 1
β
x/L
level T1
level T2
level T3
-1.5
-1
-0.5
0
0.5
1
0 0.25 0.5 0.75 1
β
x/L
level T1
level T2
level T3
50
Fig. 3.3 shows the horizontal stress distribution at different depth of transfer
beam for different models. For both of these three models, the result shows that at the
bottom level of the transfer beam, the horizontal stress distribution is in tension except
at and near the exterior and interior support, where the horizontal stress is in
compression. Besides, at this level the transfer beam subject the large variation of
horizontal stress. Moreover, model B-2 and model B-3 shows that the horizontal stress
at interior support is much larger than the horizontal stress at the end support.
However, at the middle level of the transfer beam, the result shows that it is in fully
tension stage which is unlike the normal beam. For the middle level of normal beam,
it usually shows zero stress so call the neutral axis. At the top level of the transfer
beam, the result shows that there are some portions of it still are in tension especially
for model B-1.
The largest horizontal compression stress occurs at the interior support of bottom
level of model B-2. However, the largest horizontal tension stress occurs at the bottom
level of model B-1. Besides, for the top level of model B-2 and model B-3, the results
show that the horizontal stress at mid span of transfer beam is subjected to tension but
not for model B-1. The horizontal stress is still in tension at mid span of model B-1 at
top level. Therefore, the neutral axis changes due to the importance of the length of
span.
51
This effect is because the shear deflection is also counted into calculating the
horizontal stress. For normal beam, there is an assumption that the shear deflection is
zero. Therefore, the theory and formulas for normal beam is not useful in this situation
because the stress will be govern by shear deflection if the beam is a deep beam. That
means for a deep beam or transfer beam in the shear wall-transfer beam system, the
theory and formulas for normal beam is no longer be available because the span/depth
ratio of those kind of beam is much larger than normal beam. Therefore, the shear
deflection effect will govern the total deflection of an element and the stress
distribution.
Figure 3.4 Horizontal stress along depth at mid span
0
0.2
0.4
0.6
0.8
1
-1.5 -1 -0.5 0 0.5
y/h
β
One span model
Two span model
Three span model
h=1.6
52
Fig. 3.4 shows the horizontal stress distribution along the depth of transfer beam
at mid span for different models. For normal beam, the neutral axis is usually at the
middle depth of the beam to separate the compression part and tension part of a beam.
However, for the transfer beam in the shear wall-transfer beam system, the neutral
axis goes up because the shear wall can be considered as a part of transfer beam.
Therefore the neutral axis would move up due to this interaction between shear wall
and transfer beam. For the result from one span model, the transfer beam is subjected
to fully tension situation which means the neutral axis is even upper than the top of
transfer beam and occurs in the shear wall. Moreover, the top reinforcement of the
transfer beam must be also considered when it is in the design phrase. For the results
from two and three span models, the neutral axis occurs within the transfer beam.
However, the location of neutral axis is even much upper than normal beam. With the
span of those models more which means the span/depth ratio is lower, the location of
neutral axis becomes lower. The details of this behavior will be discussed in Chapter
4.
The present of the horizontal stress distribution curve is because the present of
shear stress. It has to balance the three equilibrium equations for each element which
means the shear stress and the horizontal stress are interacted with each other. To
satisfy the vertical equilibrium, the shear stress must be toward to upper and the
53
horizontal stress must be in tensile stress. Therefore, these two will induce the
opposite moment considering the lower part of the element. As result, the larger shear
stress will induce the larger horizontal tensile stress. The shear stress will be discussed
more detail in following section.
3. Shear stress
In this section, the shear stress distribution in the lower part of shear wall is
investigated. Three different models have been investigated and for each model, the
data of shear stress have taken from three horizontal sections where are y/L=0,
y/L=0.5 and y/L=1 (L is the span length of the model, in these three models, L=15m
and y is the height of shear wall counted from the bottom of shear wall) to see the
variation of shear stress distribution at these three different locations. Those levels are
the same as the levels investigated for vertical stress.
The variation of shear stress distributions along the height of shear wall of these
three models are shown in Fig. 3.5.
54
(a) Variation of shear stress along the shear wall (Model B-1)
(b) Variation of shear stress along the shear wall (Model B-2)
-1.5
-1
-0.5
0
0.5
1
1.5
0 0.25 0.5 0.75 1
τ
x/L
level S1
level S2
level S3
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
0 0.25 0.5 0.75 1
τ
x/L
level S1
level S2
level S3
55
(c) Variation of shear stress along the shear wall (Model B-3)
Figure 3.5 Shear stress for set B models
Fig. 3.5 shows the shear stress distribution for three different models at different
location in the lower part of shear wall. For both of these models, the shear stress
distribution at the level S3 is zero that means no shear effect at or above this level
which is the same as the vertical stress part. With the location becomes lower, the
shear stress distribution becomes a curve as like those curves at location of level S2
shown in the figure. At this level the shape of shear stress distribution for these models
is most likely the same. However, at the location of level S1, the shear stress
distribution becomes different for these three models due to the effect of number of
span.
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0 0.25 0.5 0.75 1
τ
x/L
level S1
level S2
level S3
56
In the model B-1 there is a region that the shear stress tends to zero which is near
the mid span. However, for model B-2 and model B-3, the span length is shorter
therefore this effect becomes a point but not a region. Moreover, those interior
columns affect the shear stress raise up again so that the shear stress increase rapidly
again.
3.3.2 Constant span length
In this section, the length for each span is fixed as 15m. For one span model, the
total length of model is 15m which is the same as the model in section 2.3.1.
Therefore, one span model will not be investigated again in this section. For two and
three span model, the overall total length is 30m and 45m, respectively. In this section,
the investigation method is exact the same as Section 3.3.1 but the dimension of
models is different to see effect of stresses when the length of each span become
longer. The length of span for these two models are shown in Table 3.1 and the other
dimensions for each partitions of models are the same as those models discussed in
Section 3.3.1. In this section, the models name as set C.
57
1. Vertical stress
In this section, the vertical stress distribution in the lower part of shear wall is
investigated with the same span length. The two models shown in Fig. 3.1(b) and Fig.
3.1(c) have been investigated. The variation of vertical stress distribution for two
models along the length of shear wall is shown in Fig. 3-6.
(a) Variation of vertical stress along the shear wall (Model C-1)
-1
0
1
2
3
4
5
6
7
0 0.25 0.5 0.75 1
α
x/L
level S1
level S2
level S3
58
(b) Variation of vertical stress along the shear wall (Model C-2)
Figure 3.6 Vertical stress for set C models
For the two span model, the vertical stress distribution is shown in Fig. 3.6(a).
The vertical stress distribution is just shown the half of the models because of the
symmetric of the model. Compare with Fig. 3.2(b) which shows the result obtained
from two span model in Section 3.3.1, the trend of the variation of the vertical stress
distribution is not much different. However, the vertical stress is approximately equal
to zero near the mid span for Fig. 3.6(a) which is because the length of span is longer.
Therefore, the arch effect becomes more significant and most of stress transferred at
the two columns support.
-1
0
1
2
3
4
5
6
7
0 0.25 0.5 0.75 1 1.25 1.5
α
x/L
level S1
level S2
level S3
59
The vertical stress parameter , which can shows the ratio of vertical stress and
the in-plane load applied on the top of the shear wall, Compare with two results from
Fig. 3.2(b) and Fig. 3.6(a), the at the exterior and interior column support for the
span length fixed model is almost double to that obtained from total length fixed.
Which means with the same thickness of shear wall and in-plane distributed load, the
vertical stress at exterior and interior columns for the span length fixed model is
almost double to the total length fixed model.
The results from three span models are shown in Fig. 3.6(b). The difference
between these two models is similar to the two span models in Fig 3.2(c). However,
the variation of vertical stress at exterior and interior support for three span models
becomes larger than the two span models.
2. Horizontal stress
For the investigation of horizontal stress, the transfer beam is investigated to see
the tensile stress along the span length. The investigation method is also the same as
Section 3.3.1. The horizontal stress distributions for the two models are shown in
below.
60
(a) Variation of horizontal stress along the transfer beam (Model C-1)
(b) Variation of horizontal stress along the transfer beam (Model C-2)
Figure 3.7 Horizontal stress for set C models
-3
-2
-1
0
1
2
3
4
0 0.25 0.5 0.75 1
β
x/L
level T1
level T2
level T3
-3
-2
-1
0
1
2
3
4
5
0 0.25 0.5 0.75 1 1.25 1.5
β
x/L
level T1
level T2
level T3
61
Fig. 3.7 shows the horizontal stress distribution at different levels of transfer
beam for two span and three span models. The horizontal stress distribution is just
shown the half of models because of symmetric of models. Compare the results of Fig.
3.7(a) and Fig. 3.3(b), which is both two span cases, the trend of horizontal stress
distribution is similar. However, the horizontal stress at the bottom of transfer beam in
Fig. 3.7(a) is more than double of Fig. 3.3(b). Besides, Fig. 3.7(a) shows most of the
part of transfer beam is subjected a tensile force even at the top of the transfer beam
except the part near supports. It can say that this transfer beam is a fully tension beam
due to the large span length. This is different from the result obtained from Fig. 3.3(c),
which shows the horizontal stress at the top of transfer beam is positive that means at
that location, the transfer beam is subjected a compressive force.
The horizontal stress at the bottom of transfer beam in Fig. 3.7(b) is much larger
than horizontal stress in Fig. 3.3(c). Besides, the transfer beam in Fig. 3.7(b) is a fully
tension beam.
62
3. Shear stress
For the shear stress investigation, the lower part of shear wall is investigated. The
investigation method for shear stress is also the same as the method used in Section
3.3.1. Fig. 3.8 shows the shear stress distribution at different levels for the lower parts
of shear wall for set C models.
(a) Variation of shear stress along shear wall (Model C-1)
-1.5
-1
-0.5
0
0.5
1
1.5
0 0.25 0.5 0.75 1
τ
x/L
level S1
level S2
level S3
63
(b) Variation of horizontal stress along shear wall (Model C-2)
Figure 3.8 Shear stress for set C models
The shear stress distributions in Fig. 3.8 are also just shown the half of models
due to symmetric. Compare with the shear stress distribution in Fig. 3.8 and Fig. 3.5,
the shear stress obtained from Fig. 3.8(a) is also large than the shear stress obtained
from Fig. 3-5(b). Besides, the shear stress distribution at the bottom of shear wall in
Fig. 3.8(a) shows there are two inflection points. Between these two inflection points,
the shear stress decrease slowly where is near the mid span of transfer beam.
Moreover, the shear stress is very small near mid-span. This is different from the shear
stress distribution at the bottom of shear wall in Fig. 3.5(b) which is because there is
-1.5
-1
-0.5
0
0.5
1
1.5
0 0.25 0.5 0.75 1 1.25 1.5
τ
x/L
level S1
level S2
level S3
64
no inflection point for the shear stress distribution near the mid span in Fig. 3.5(b).
Therefore, the shear stress decreases rapidly even near the mid span of model. For the
shear stress at the level S1 in Fig. 3.8(a), there is still some shear stress near the mid
span. However, the shear stress at that location is small. Besides, the trend of shear
stress distribution at the level S2 is also similar compare to the results from Fig. 3.8(a)
and Fig. 3.5(b).
The shear stress distribution shown in Fig. 3.8(b) Fig. 3.5(c), there are also
similar to the two span cases.
2.3 Summary
In this Chapter, the effect of structural behavior on a shear wall-transfer beam
system due to span which including the number of span and the length of span is
investigated. The variation of stresses distribution in a shear wall-transfer beam
system can show the structural behavior. Therefore, the vertical stress and shear stress
in the lower part of shear wall and the horizontal stress in the transfer beam at
different levels are taken out from different models and compared the changes due to
different conditions in this Chapter.
65
1. Vertical stress
For vertical stress at the lower part of shear wall in a shear wall-transfer beam
system, the main investigating direction is the arch effect behavior which is the
behavior that the vertical stress near the mid span of the model would transmit to the
two supporting columns. Because of this effect, the vertical stress near mid span
becomes small and the vertical stress at the two supporting columns become large.
The results show that the significance of arch effect depends on two factors which are
the number of span and the length of span.
The number of arch depends on the number of span. If there is one span model,
there is only one arch for the vertical stress and so on. Besides, the vertical stress at
the exterior column is much larger than the vertical stress at the interior column.
On the other side, the length of span decides the shape of arch. If the length of
span is long enough, the vertical stress near mid span would transmit to the two
nearest columns and tends to zero. That means the larger length of span, the larger
vertical stress at two columns as result when all other factors are in the same scenario.
However, if the length of span is short, the vertical stress would increase quickly from
the mid span. That results a less vertical stress at the two columns because those
vertical stress near mid span has no enough space to transmit to two columns.
Therefore, some of vertical stress would stay near mid span.
66
2. Horizontal stress
For the horizontal stress at transfer beam in a shear wall-transfer beam system,
the stress shows the structural behavior of transfer beam whether it subjects a tensile
stress or compression stress. In this Chapter, the effect of horizontal stress due to
number of span and length of span are investigated. Besides, the non-linear effect on
horizontal stress in transfer beam is shown. However, this effect is more related to
span/depth ratio. Therefore, the more details for this effect will be discussed in
Chapter 4.
From the results shown in this Chapter, at the level T1, the horizontal stress
increase and become positive rapidly at the regions where there are column supports.
That means the horizontal stress changes from tensile stress to compressive stress
rapidly near the column support. However, this effect reverses for the level T2 and
level T3. From this result, it can say that more the number of span would result more
curves in the horizontal stress distribution. Besides, the horizontal stress at interior
column support is much larger than the horizontal stress at exterior column.
Meanwhile, the shape of horizontal stress distribution near mid span is controlled
by the length of span. If the length of span is long enough, the horizontal stress
distribution near mid span would become more flat as result.
67
3. Shear stress
For the shear stress at the lower part of shear wall in a shear wall-transfer beam
system, the two factors which are number of span and length of span are investigated
that how would affect the structural behavior in a shear wall-transfer beam system.
From the results in this Chapter, the shear stress distribution would pass through
the zero point when there is an interior column. Besides, at the level S1, the shear
stress increases rapidly if there is a column support. This is the effect that affect by the
number of span. However, for the other levels in the lower part of shear wall, this
conclusion is not available.
On the other hand, the length of span controls the shape of shear stress
distribution near mid span. At the level S1, there are two inflection points near mid
span when the length of span is long enough. Between these two inflection points, the
shear stress distribution becomes less variation and the shear stress close to zero.
However, these two inflection points become closer when the length of span becomes
shorter.
68
CHAPTER 4 EFFECT OF SPAN/DEPTH RATIO
4.1 Introduction
Besides of the structural behavior of shear wall-transfer beam system due to
number of span and span length, the other main factor that would affect the structural
behavior of shear wall-transfer beam system is the span/depth ratio. In this Chapter,
the effect of stress distribution due to different span/depth ratio will be investigated.
Usually the transfer beam in shear wall-transfer beam system is much deeper than
normal beam therefore the transfer beam can sustain the high moment and shear force
induced by heavy load which included the self weight of shear wall and the load
adding on the shear wall.
In this Chapter, two sets of model, which are constructed to investigate the effect
of structural behavior due to span/depth ratio, name as set D and E. Each set of model
contains three models which are one span shear wall-transfer beam structures but with
different span/depth ratio. Another set of model contains also three models which are
two span shear wall-transfer beam structures but with different span/depth ratio.
These two sets of models are simulated when the top of shear wall subject an
in-plane pressure P toward to transfer beam where P equal to 100kN/m2. Therefore,
with the thickness of shear wall 0.4m, the distributed in-plane load equal to 40kN/m.
69
For this six models, the length of span is fixed but with different depth of transfer
beam to change the span/depth ratio. Table 4.1 shows the span/depth ratio for these
models. Fig. 4.1 and Fig. 4.2 show the geometry of all models.
Table 4.1 Span/depth ratio for models
Span/depth ratio
One span Two span
Model 1 15 15
Model 2 10 10
Model 3 7.5 7.5
(a) Model with one span (Model D-1)
70
(b) Model with one span (Model D-2)
(c) Model with one span (Model D-3)
Figure 4.1 One span models with span/depth ratios
71
(a) Model with two span (Model E-1)
(b) Model with two span (Model E-2)
72
(c) Model with two span (Model E-3)
Figure 4.2 Two span models with span/depth ratios
4.2 Structural behavior
The investigation method in this Chapter is similar to the investigation method in
Chapter 3. However, the investigation factor is not the same. The main difference
between a normal beam and deep transfer beam is that the shear effect is very small
for a normal beam. Therefore, the shear strain is ignored in simulate the structural
behavior for a normal beam and the main factor for total strain is bending. However, a
beam becomes deeper, the shear effect becomes larger and the bending effect becomes
not the main factor for total strain but shear effect. For a transfer beam in shear
wall-transfer beam system, the beam is usually very deep because the transfer beam is
73
supporting a high vertical loading which including the self weight of shear wall and
the loading on each floor. Thus, the shear effect must be taken account in simulating a
transfer beam in shear wall-transfer beam system.
From the results in Chapter 3, it shows that the vertical stress and shear stress at
the top of lower part of shear wall becomes constant and zero, respectively. Therefore,
the vertical stress and shear stress will not be investigated at the level S3 in this
Chapter. The vertical stress and shear stress will be investigated at the level S1 and
level S2. Besides, the horizontal stress will be investigated at three levels which are
level T1, level T2 and level T3.
4.3 Results and analysis
The set of models will be investigated and compare the results with different
span/depth ratio. The set of models will be separated in to two parts, one span models
and two span models for investigation. For each part, three stresses which including
vertical stress, horizontal stress and shear stress will be investigated and for each
stress. For vertical stress and shear stress in the lower part of shear wall, two levels of
that partition will be investigated and for horizontal stress in the transfer beam, three
levels will be investigated.
74
4.3.1 One span models
For one span investigation, there are three models that would be investigated to
compare the variation of stresses due to different span/depth ratio.
1. Vertical stress
Below shows the vertical stress distribution for three different models at level S1
and level S2.
(a) Variation of vertical stress at level S1
-2
0
2
4
6
0 0.2 0.4 0.6 0.8 1
α
x/L
L/d=15
L/d=10
L/d=7.5
75
(b) Variation of vertical stress at level S2
Figure 4.3 Vertical stress for set D models
Fig. 4.3 shows the vertical stress distribution at different levels for three one span
models with different span/depth ratio. Fig. 4.3(a) shows that at the column support,
the vertical stress becomes larger when the span/depth ratio becomes larger. However,
the vertical stress drops quicker if the span/depth ratio becomes larger. Therefore,
Model 1 which has the larger span/depth ratio results the larger vertical stress at the
column support but the vertical stress from model 3 exceed the vertical stress from
model 1 quickly when there is no column support. Nevertheless, the vertical stresses
from three models tend to zero near mid span. That is because the length of span is
long and the arch effect for these models become significant. In order to compare the
0.5
1
1.5
0 0.2 0.4 0.6 0.8 1
α
x/L
L/d=15
L/d=10
L/d=7.5
76
variation of vertical stress near mid span, the vertical stress distribution at level S2 for
three models is shown in Fig. 4.3(b). From the results in Fig. 4.3(b), the trend of
vertical stress distribution for three models is almost the same as the results from Fig.
4.3(a) except near the mid span. The arch effect for vertical stress at level S2 is much
less then at level S1. Therefore, the decrement of vertical stress is not fast for three
models. Besides, the larger span/depth ratio results the larger vertical stress at mid
span.
2. Horizontal stress
Below shows the horizontal stress distribution for three different models at level
T1, level T2 and level T3.
77
(a) Variation of horizontal stress at level T1
(b) Variation of horizontal stress at level T2
-3
-2
-1
0
1
2
0 0.2 0.4 0.6 0.8 1
β
x/L
L/d=15
L/d=10
L/d=7.5
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0 0.2 0.4 0.6 0.8 1
β
x/L
L/d=15
L/d=10
L/d=7.5
78
(c) Variation of horizontal stress at level T3
Figure 4.4 Horizontal stress for set D models
Fig. 4.4 shows the horizontal stress distribution for different models at level T1,
level T2 and level T3 respectively. It is observed that the larger span/depth ratio which
would result the larger horizontal tensile stress in most part of transfer beam. Fig.
4.4(c) shows the horizontal stress at the top of the beam is in compression if the
span/depth ratio is not large. However, the transfer beam can be said that which is a
fully tension beam when the span/depth ratio is large because the horizontal stress
distribution at the top of beam shown in Fig. 4.4(c) is in tension at most part of the
beam.
-1
-0.5
0
0.5
0 0.2 0.4 0.6 0.8 1
β
x/L
L/d=15
L/d=10
L/d=7.5
79
3. Shear stress
Below shows the Shear stress distribution for three different models at level S1
and level S2.
(a) Variation of shear stress at level S1
(a) Variation of shear stress at level S2
Figure 4.5 Shear stress for set D models
-1.5
-1
-0.5
0
0.5
1
1.5
0 0.2 0.4 0.6 0.8 1
τ
x/L
L/d=15
L/d=10
L/d=7.5
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0 0.2 0.4 0.6 0.8 1
τ
x/L
L/d=15
L/d=10
L/d=7.5
80
Fig. 4.6 shows the shear stress distribution for different models at level S1 and
level S2 respectively. From Fig. 4.6(a), the shear stress increases quicker when the
span/depth ratio becomes larger. However, after the peak of shear stress occurs, the
shear stress drops also quicker if the span/depth ratio is larger. Besides, it is observed
that when the span/depth ratio is larger, there are more portions near mid span that the
shear stress closes to zero. For the shear stress shown in Fig. 4.6(b), it is observed that
the shear stress is always larger if the span/depth ratio becomes larger.
4.3.2 Two span models
In this section, the three stresses obtained from three different models with
different span/depth ratio will be compared with those one span models in Section
4.3.1.
1. Vertical stress
Below shows the vertical stress distribution for different models at level S1 and
level S2.
81
(a) Variation of vertical stress at level S1
(b) Variation of vertical stress at level S2
Figure 4.6 Vertical stress for set E models
-1
1
3
5
7
0 0.5 1 1.5 2
α
x/L
L/d=15
L/d=10
L/d=7.5
0
0.5
1
1.5
2
0 0.5 1 1.5 2
α
x/L
L/d=15
L/d=10
L/d=7.5
82
Fig. 4.7 shows the vertical stress distribution at different levels in the lower part
of shear wall. The trends of vertical stress at each span are similar to one span case.
However, the vertical stress increases quickly near the interior support and the vertical
stress at interior support becomes larger when the span/depth ratio becomes larger.
2. Horizontal stress
Below shows the horizontal stress distribution for different models at level T1,
level T2 and level T3.
(a) Variation of horizontal stress at level T1
-3
-2
-1
0
1
2
3
0 0.5 1 1.5 2
β
x/L
L/d=15
L/d=10
L/d=7.5
83
(b) Variation of horizontal stress at level T2
(c) Variation of horizontal stress at level T3
Figure 4.7 Horizontal stress for set E models
-1.5
-1
-0.5
0
0.5
1
0 0.5 1 1.5 2
β
x/L
L/d=15
L/d=10
L/d=7.5
-1.5
-1
-0.5
0
0.5
1
0 0.5 1 1.5 2
β
x/L
L/d=15
L/d=10
L/d=7.5
84
Fig. 4.8 shows the horizontal stress distribution for different models at different
levels. Compare with Fig. 4.4 which is the horizontal stress distribution for one span
model, the trends of distribution are also similar except near the interior column. At
the level T1, the horizontal stress changes to compression quickly near the interior
column. Besides, at the center line of interior column the horizontal stress drops down
a little bit and this effect becomes more significant when the level of transfer beam
becomes higher.
3. Shear stress
Below shows the shear stress distribution for different models at level S1 and
level S2.
(a) Variation of shear stress at level S1
-1.5
-1
-0.5
0
0.5
1
1.5
0 0.5 1 1.5 2
τ
x/L
L/d=15
L/d=10
L/d=7.5
85
(b) Variation of shear stress at level S2
Figure 4.8 Shear stress for set E models
Fig. 4.9 shows the shear stress distribution at the lower part of shear wall for
different models at different levels. It shows that the trend of shear stress for two span
models and one span model is also similar. The different of shear stress distribution
between two span is just midpoint symmetry.
-0.4
-0.2
0
0.2
0.4
0 0.5 1 1.5 2
τ
x/L
L/d=15
L/d=10
L/d=7.5
86
4.3.3 Non-linear stress distribution effect in transfer beam
From Chapter 3, one of behavior of the transfer beam in shear wall-transfer beam
system is that the distribution of horizontal stress along the depth of transfer beam
becomes non-linear when the span of model becomes more, or says the span/depth
ratio becomes lower. Fig. 3.4 shows the horizontal stress distribution along the depth
of transfer beam is still linear for one span situation. However, the non-linear behavior
of stress distribution occurs from the two span model and more significant in the three
span model.
To investigate this kind of effect, a set of model have been generated to perform
the finite element analysis. Consider five beams with different span/depth ratio but the
depth of those beams is fixed as 2m. That means the span length various to control the
span/depth ratio. The boundary condition for all of beams is simply supported and
supporting the in-plane distribute load at the top face of the beam which magnitude is
100kN/m and the thickness of beam is 1m. The results show as below:
87
Figure 4.9 Horizontal stress for models
Fig. 4.10 shows the horizontal stress distribution for different models. From the
L/d=0.5 model, the shape of horizontal stress distribution shows the linear distribution
which is as like as a normal beam and the location of neutral axis is about the mid
depth of the beam. However, when the span length/depth ratio becomes higher, the
distribution becomes non-linear and the values of horizontal stress become lower for
both compression region and tensile region. That means for a deep beam, the upper
part of the beam will subject a smaller horizontal stress and tends to zero if the beam
is deep enough. The stress will induce at the lower part of beam only therefore the
location of neutral axis will become lower.
0
0.2
0.4
0.6
0.8
1
1.2
-10 -5 0 5 10
d/h
β
L/d=3
L/d=2
L/d=1
L/d=0.5
L/d=0.25
88
4.4 Summary
In this Chapter, the effect of stresses for one span and two span models due to
span/depth ratio is investigated. Besides, the non-linear horizontal stress distribution
effect in transfer beam due to span/depth ratio is also investigated.
1. Vertical stress
The results show that the vertical stress becomes larger when the span/depth ratio
becomes larger at the column supports. However, the increase and decrease of vertical
stress is also quicker when the span/depth ratio becomes larger. The different of trend
for vertical stress between one span model and two span model is just there are sudden
change at the interior column support for two span model.
2. Horizontal stress
The results show that the horizontal tensile stress becomes larger when the
span/depth ratio becomes larger near the mid span. Compare to the one span and two
span cases, the result shows that there is sudden change for horizontal stress near the
interior support for two span models. Besides, the variation of horizontal stress due to
span/depth ratio becomes smaller.
3. Shear stress
The results show that the shear stress becomes larger when the span/depth ratio
becomes larger. However, the shear stress near mid span is close to zero at level S1.
89
Compare to the one span and two span models, the trend of shear stress distribution is
also similar for different levels. Besides, the shear stress distribution is point
symmetric at the midpoint of model for two span models.
4. Non-linear stress distribution effect in transfer beam
The horizontal stress distribution along the depth of transfer beam at mid span
becomes non-linear when the span/depth ratio becomes smaller which means the
beam becomes thicker. Besides, when the span/depth ratio becomes smaller, the upper
part of beam is useless because the horizontal stress at the upper part of beam is zero.
The compressive stress and tensile stress are presented at the lower part of the beam
and those stresses are small.
90
CHAPTER 5 CONCLUSION
In this paper, three parameters, which are the number of span, the length of span
and the span/depth ratio, would affect the structural behavior of a shear wall-transfer
beam system are investigated.
5.1 Effects on structural behavior
The number of span decides the number of sudden change in stress. Usually the
stress changes reversely because there is support which means the more boundary
condition for the near elements. Besides, the length of span decides the shape of
stresses. The longer length of span, the stresses would keep increasing or decreasing
further until there is another column support. On the other hand, the span/depth ratio
decides the also the trend of stresses and the magnitude of the stresses. Usually the
stresses would be larger when the span/depth ratio becomes larger.
5.2 Recommendation for future study
After the investigation for the shear wall-transfer beam system based on the
effect of span and the span/depth ratio, there are several directions for the future study
about this topic:
91
1. Effect on loading
In this paper, the shear wall transfer beam system is subjected to an in-plane load
at the top of shear wall. However, besides of in-plane load there is a lateral load like
wind load which is also subjected by shear wall. The structural behavior of shear wall
transfer beam system would be totally different if there is a lateral load adding on the
shear wall.
2. Effect by other geometry parameter
The geometry of the shear wall-transfer beam system is one of the main factors
that would affect the structural behavior. Besides of the number of span, length of
span and the span/depth ratio, there are some of geometric parameters which can be
investigated like the thickness ratio of shear wall and transfer beam, and the size of
column.
3. Opening on shear wall
For a modern high-rise building, there are some opening on the shear wall to
provide facade or ventilation area. However, the structural behavior of shear
wall-transfer beam system would be affected much when there is any open area on the
shear wall. Therefore, this is also one of direction for studying the effect of structural
behavior due to open are on shear wall.
92
References
Jawaharlal, P. (1996), Analysis and behavior of transfer girder system in tall buildings,
PhD Thesis, The Hong Kong University of Science and Technology, Hong Kong.
Kuang, J.S. and Jawaharlal, P. (1996), “Continuous transfer beams supporting in-plane
loaded shear walls in tall buildings”, The Structural Design of Tall Buildings, Vol.5,
pp. 281-293.
Mir M. Ali and Kyoung, S. M. (2007), “Structural developments in tall buildings:
current trends and future prospects”, Architectural Science Review, Vol.50, No 3, pp.
205-223.
Yettram, A.L. and Hirst, M.S. (1971), “An elastic analysis for the composite action of
walls supported on simple beams”, Buildings Sci., Vol.6, pp. 251-159.
Smith, B.S. and Riddington, J.R. (1977), “The composite behavior of elastic
wall-beam systems”, Proc. Instn Civil Engrs, Vol.63, No 2, pp. 377-391.